The Control System of CERN Accelerators Vacuum (LS1 Activities

CERN-ACC-2014-0352
Paulo.Gomes@cern.ch
The Control System
of CERN Accelerators Vacuum
(LS1 Activities and New Developments)
P. Gomes, F. Antoniotti, F. Bellorini, S. Blanchard, J.P. Boivin, J. Gama, G. Girardot, G. Pigny, B.
Rio, H. Vestergard
CERN, Geneva, Switzerland
L. Kopylov, S. Merker, M. Mikheev
IHEP, Protvino, Russia
17/12/2014
CERN-ACC-2014-0352
Keywords: LS1, Vacuum
Abstract
After 3 years of operation, the LHC entered its first Long Shutdown period (LS1), in February 2013
[1]. Major consolidation and maintenance works are being performed across the whole CERN’s
Accelerator chain, in order to prepare the LHC to restart at higher energy, in 2015. The injector chain
shall resume earlier, in mid-14.
We report about the on-going vacuum-controls projects. Some of them concern the renovation
of the controls of certain machines; others are associated with the consolidations of the vacuum
systems of LHC and its injectors; and a few are completely new installations.
ue to the wide agespan of the existing vacuum installations, there is a mix of design philosophies and of controlequipment generations. The renovations and the novel projects offer an opportunity to improve the
uniformity and efficiency of vacuum controls by: reducing the number of equipment versions with
similar functionality; identifying, naming, labelling, and documenting all pieces of equipment;
homogenizing the control architectures, while converging to a common software framework.
Presented at ICALEPCS 2013, 6-11 October 2013, San Francisco, California
Geneva, Switzerland
November, 2014
EDMS no 1335717
THE CONTROL SYSTEM OF CERN ACCELERATORS VACUUM
[ LS1 ACTIVITIES AND NEW DEVELOPMENTS ]
P. Gomes, F. Antoniotti, F. Bellorini, S. Blanchard, J.P. Boivin,
J. Gama, G. Girardot, G. Pigny, B. Rio, H. Vestergard
CERN, Geneva, Switzerland
L. Kopylov, S. Merker, M. Mikheev ; IHEP, Protvino, Russia
Abstract
PS & AD RENOVATIONS
After 3 years of operation, the LHC entered its first
Long Shutdown period (LS1), in February 2013 [1].
Major consolidation and maintenance works are being
performed across the whole CERN’s accelerator chain, in
order to prepare the LHC to restart at higher energy, in
2015. The injector chain shall resume earlier, in mid-14.
We report about the on-going vacuum-controls
projects. Some of them concern the renovation of the
controls of certain machines; others are associated with
the consolidations of the vacuum systems of LHC and its
injectors; and a few are completely new installations.
Due to the wide age-span of the existing vacuum
installations, there is a mix of design philosophies and of
control-equipment generations. The renovations and the
novel projects offer an opportunity to improve the
uniformity and efficiency of vacuum controls by:
reducing the number of equipment versions with similar
functionality; identifying, naming, labeling, and documenting all pieces of equipment; homogenizing the
control architectures, while converging to a common
software framework.
ARCHITECTURES
In the 90's, the vacuum controls of the PS-Complex and
SPS were renovated to an architecture based on CPUs in
VME crates (Versa Module European), called DSC
(Device Stub Controller); these were remotely accessible
via Ethernet on a graphical interface, firstly based on XWindows, and later on Java “Working-Sets”.
The DSC collected data through an RS232/X25
interface: either directly, from industrial controllers
equipped with RS232; or from the custom-made frontends, through G64 concentrators.
In the years 2000, the SPS was upgraded to a PLC
(Programmable Logic Controller)-based architecture with
Siemens®-S7 series. It was followed by the new LHC and
by the partial renovation of the PS-Complex. The PLC
accesses field equipment: directly on a fieldbus
(Profibus®) or via remote-IO stations (Siemens ET200).
The Supervisory Control And Data Acquisition
(SCADA) is built on PVSS® (Siemens WinCC-OA); its
Data Servers (PVSS-DS) talk with the PLCs via Ethernet.
Both PLC & SCADA follow the UNICOS framework [2].
VGR/P
PS
AD
86
72
VGI
19
VPG
VPS
VPI
16
117
100
156
65
VPC
VVS
TOT
6
13
15
372
293
Table 1: instrument count for the PS and AD renovations
The Proton Synchrotron (PS) is part of the LHC
injector chain; the Antiproton Decelerator (AD) produces
low-energy antiprotons for the study of antimatter. Within
the CERN accelerator chain, only the PS and AD kept
their legacy control system until today.
During LS1, a significant amount of maintenance and
consolidation will be performed on both PS & AD
machines, in order to improve their reliability for the next
20 years. This is an occasion for renovating their vacuum
control system, by propagating the PLC/PVSS architecture, bringing the hardware & software to the level of
the other machines, and thus to enforce standardization.
The TPG300® controllers, for Pirani/Penning gauges
(VGR/P), need simply to have their communication card
replaced from RS232 to Profibus. The old custom controllers for Bayard-Alpert gauges (VGI) cannot be upgraded
and will be replaced by modern Volotek ®, already with a
Profibus interface.
The old ion-pump (VPI) controllers will be replaced by
recent ones, accessed over standard remote-IO stations.
The sublimation-pump (VPS) controllers will be kept,
connected through newly-designed remote-IO stations.
That will also be the case for the cryogenic-pumps (VPC).
In AD, the outdated pumping-group (VPG) controllers
will be eliminated or replaced by standard models.
Profibus will be available for the mobile groups, along the
AD ring. In the PS there will be no renovation of VPG.
The valve (VVS) controllers will be upgraded, and
controlled directly from the PLC local-IO.
The power distribution to the racks will be redimensioned and standardized. The most critical
components (such as PLC, valve & interlock controllers,
and some pumping-groups) will be backed-up by UPS at
AD and by Diesel-generator in PS.
A power line will be pulled throughout the AD ring, for
the heaters mounted on the VPI electrical feed-through, to
prevent humidity and oxidation.
All cable numbers and connections will be identified, in
order to revise the electrical documentation; in particular,
the logic and wiring of the interlocks (for pumps, valves,
and client systems) will be checked and updated. Once
the new controls are fully commissioned, the old
components will be definitely removed and discarded.
Like for the other machines in the CERN accelerator
complex, Working-Sets will be replaced by PVSS. New
SCADA and PLC applications will be developed for the
specific controls of VPS and VPC.
CPS CONSOLIDATIONS
TT10 consolidation
Apart from the renovations of PS & AD, maintenance
and consolidation works will be applied to the other
machines in the PS-Complex. This is also the occasion to
replace the damaged cables.
LINAC2
The first 3 pumping groups, near the source,
already been upgraded to connect to Profibus;
remaining 6, on the accelerating cavities, will
become remotely accessible, without requiring
presence in the rack room to be operated.
had
the
also
the
The injection line into the SPS will have its instruments
renovated. While the control of the gauges and valves was
already of the modern flavor, new cables and new
standard controllers are needed for the VPIs. The 4 racks
will be re-arranged by vacuum sector.
Racks re-organization
Renovated a few years ago, the Booster will now have
its 43 VPS remotely controlled, like PS & AD. Indeed,
never before the VPS had been integrated in the Profibus/
PLC architecture of any machine; the remote-IO controller and associated software will be developed for LS1.
With the successive addition of new equipment along
the years, re-structuring the SPS racks topology became
necessary. We will start with the ones already concerned
by current activities: TT10 (4 racks), and the ARCs (6x4).
Besides, the full 15 racks in BA5, and their internal
cabling, will also be organized by sector. The controllers
will thus be better recognizable and accessible; with some
slots reserved for expansion on each rack, it will be easier
to add new equipment to any sector, without breaking the
order. This prototype work will inspire the evaluation of
feasibility and effort needed for the rest of the machine.
Profibus for mobile
Damaged Cables
In LEIR, a temporary Profibus network will be
installed, to provide remote access to mobile devices,
such as the “bake-out” controllers. The installation of a
Profibus network for mobile devices, in PS, PSB, and TLs
(Transfer Lines), was initially thought for the LS1; this
was postponed due to budgetary constraints.
The radiation accumulated by the cables in the tunnel
progressively degrades their characteristics, eventually
destroying them. As part of a campaign of cyclic renovation, all the cables on the right of point1 (TS1+) will be
replaced; likewise for the TDC2/TCC2 target area.
In the transfer line to the North Area (TT20), only the
damaged cables will be replaced. In left of point1 (TS1-),
the number of problems was steadily increasing; all
cables will be replaced. A patch-panel rack at the bottom
of the pit will be used to split the segment in the tunnel
from the one up to the racks.
PS Booster
TLs
All devices in the Transfer Lines are presently
controlled by the PLCs of LINAC2, PS & PSB. When
LINAC4 will replace LINAC2, the controls should all be
concentrated on the PLC of PSB. Technically possible in
LS1, this was also postponed.
SPS CONSOLIDATIONS
New sectorization
Each of the 6 ARCs in the SPS is divided in only 2
vacuum sectors, of about 400 m. To reduce the pumping
time for these long volumes, they will both be split into 2.
Therefore, the 12 new sectors will need to be equipped
with extra instruments, cables, and controllers.
Due to budgetary and manpower limitations, the installation of cables and instruments will not take place during
LS1. However, the control racks are being prepared in
advance: the new controllers will be installed; the 4 racks,
corresponding to the 4 new sectors per Sextant, will be rearranged, together with their internal cables.
In point1 (BA1), additional sectors will be created
around a Kicker and a Dump, isolating them from leaks in
the vicinity. Apart from the usual extra instruments, a
special pendulum valve has to be integrated.
In point5 (BA5) two sector valves, which have been in
local mode for some time, will be integrated in the control
system, therefore inducing the creation of 2 new sectors.
Also in BA5, new gauges will be added to assess the
performance of Carbon coating on beam pipe in magnets.
LHC CONSOLIDATIONS
Several LHC subsystems will require new instruments
or consolidations. NEG cartridges will be added to
increase pumping speed on Inner Triplets, Stand Alone
Magnets, collimator areas, VAX areas, and ADT in IP4;
the corresponding power supplies will be integrated in the
control system, to perform the activations remotely.
New instruments are also needed for NEG pilot-sectors
and electron-cloud diagnosis. Remotely monitored thermocouples will be installed near the collimators (IP3, 7).
Some 570 new control cables are required all around
the LHC. The preparation and check of the cabling
requests has been done. The advancement of the ongoing
cabling campaign determines the schedule for the
equipment installation and commissioning.
Pumping groups
One additional fixed pumping group (VPGF) will be
installed at each extremity of the QRL, in order to provide
pumping redundancy; this implies 16 new controllers and
the associated cables.
Given the number of chambers opened during LS1,
extra pumping capacity will be needed: 75 old mobile
pumping groups (VPGM) will be renewed, and their controllers upgraded with touch-panels & Profibus interface.
New PLC software will be deployed for all VPGF
controllers, based on a state machine sequencer, and
incorporating auto-restart & auto-venting.
Bypass valves
For the ARC magnets, each insulation vacuum sector
should be connectable to an adjacent sector, to use its
pumping as backup if needed. Therefore, every 1 out of 2
vacuum barriers must have a bypass valve (total of 7 per
LHC ARC). Although the cabling was ready, those valves
were not initially mounted in 3/8 LHC sectors. They will
now be installed and integrated in the control system.
New interlocks
For earlier closure, the beam sector valves at the ARC
extremities will trigger on a pressure rise on Q12/Q13
instead of Q7; this requires 32 new gauge pairs VGR/P
and the associated cables and controllers.
On the other hand, a re-organization of the VPI
interlocks around IP3 & IP7 has been postponed.
Mobile equipment
New bake-out controller racks were recently produced
(20+25), with additional software functionalities [3].
The amount of mobile units (pumping-groups & bakeout racks), with predefined Profibus addresses, has been
increasing. As the number of possible addresses is
limited, they will have to be re-distributed; the definition
of device-sets per machine might become mandatory.
The temporary connections, of mobile devices to the
Profibus network in the tunnel, do not always respect the
basic rules of linear topology and line termination. To
avoid the degradation or loss of communication, often
arising in crowded zones, a consolidation of the network
is essential, starting with fixed connection boxes to avoid
star-points. The training of the users is a crucial point.
R2E
The level of radiation expected from the collimation
system at IP7 will certainly induce perturbations on
standard electronics. Therefore, all vacuum controls
equipment in UJ76 will be moved into a more protected
zone, further away from the accelerator: 27 racks will be
relocated; 282 cables must be extended from 300 to
500 m, using patch-panels. Special care must be taken
with the handling, and routing into the racks, of such a
large quantity of cables & connectors [4].
Moreover, the vacuum in the ARCs is measured by
active gauges (with local electronics), subject to radiation.
In order to better understand their tolerance, irradiation
tests are being prepared. In parallel, a new and more
accurate active gauge is under evaluation; it will also be
irradiated to evaluate its expected life time.
NEW PROJECTS
HIE-ISOLDE
Like ISOLDE before, its extension will also be based
on the UNICOS framework; it will be installed in 201418. The main challenge will be the cryogenic modules,
where beam & insulation vacuums are merged.
LINAC4
The new linear accelerator, at the start of the LHC
injector chain, will be installed in 2013-15. The control
architecture is the same as LHC, with some specific developments, like the hydrogen gas-injection for the source.
NA62
The renovation of the experimental line and detectors
will be performed during 2014. The main challenge is the
mix of turbo-molecular pumps and large cryogenicpumps, for very wide beam pipes & detector chambers.
nTOF
In 2014, during the renovation of this short secondary
beam experimental line, the vacuum controls will be
integrated within the PS-Complex PLC/SCADA.
HIE-IS
LIN4
NA62
nTOF
VGR/P VPP TMP VPI/VPC NEG/EH/TT VVS TOT
68+5 10 28 VPC: 1
20 132
40
19 17 VPI: 50 NEG: 8
8 142
48
19
8 VPC: 8 EH/TT: 14
5 102
12
5
2
19
Table 2: instrument count for the new projects
QUALITY MANAGEMENT
In order to improve the efficiency of the services
provided by the controls team, to vacuum experts and to
accelerator operators, a Quality Management Plan [5] is
being put into place. Several methods & tools have been
defined, and are progressively being implemented & used.
Naming
The vacuum controls architecture and equipment have a
history rich of a few decades; several versions of the same
equipment, from different generations, may coexist even
in the same machine. Inspired from the naming usage in
the LHC, and in conformity with the “LHC QA Definition”, we have set the coding rules for vacuum controls
equipment, independently of the machine.
VTL
To optimize resources and time, the traceability of
problems, requests, repairs, and other actions, has been
put into place. The Vacuum-controls Tracking Log (VTL)
provides a means to: follow the dates of request,
execution and finish; assign a level of urgency; record the
solutions implemented; highlight the response delay, the
engaged resources and the volume of handled issues.
MTF
Versions & UNICOS Convergence
Extensive effort is being put on the identification and
labeling of each individual device, with a coded “partidentifier”. Together with the registration in a central
database (MTF), this will allow to trace its history and
thus: optimize future repairs; organize preventive maintenance actions; manage versions, spares and obsolescence.
The PVSS package has been evolving, from ActiveX to
QT graphics and from V3.6 to V3.8, on Linux; it will
soon go to V3.11, at last fully compatible with Win-7.
History archiving, limited to about one year deep, will be
moved to an Oracle server, independently maintained.
While preserving the current functionalities, a convergence towards the CERN UNICOS framework is under
preparation; new libraries of objects & functions, tailored
for vacuum, will be developed for both PLC and SCADA.
Having it ready, for Linac4 in 2015, will be the first fullscale experience, before global deployment through the
whole accelerator chain during LS2.
EDMS
An important ongoing effort is the gathering of old
documents and the centralization of information on architectures, procedures, equipment parameters & settings, in
an Oracle database (EDMS). It will allow keeping
information up-to-date, coherent, sharable, and reliable.
CONCLUSIONS
Layout
The topology of all instruments & control devices will
be described in Layout-DB, the source for control systems
configuration and for on-line operational information.
SCADA DEVELOPMENTS
The SCADA applications for vacuum have been under
significant evolution, regarding their ergonomics,
configurability and standardization [6]:
• simplified and normalized presentation of information;
• coherent functionalities & menus, across all machines;
• automatic scripts, instead of fastidious manual actions;
• enhanced tools, for data analysis and interventions.
The access to the interlock thresholds stored in the TPG
has been improved; all TPGs on a given PLC can now be
seen simultaneously; the tracking of parameters was
extended to machines beyond the LHC.
The main SCADA panel of each machine shows now
the status of every sector valve, represented by colorcoded circles. The renovated Replay function animates
the past evolution of synoptics and bar-graphs, with the
purpose of analysis and education.
Other tools, for data-mining & data-analysis, are currently being studied, in order to be operational in 2015.
The PLC Remote-Reset is integrated in the SCADA.
Notifications
The mail/SMS notifications have been enhanced,
driven by the increasing amount and variety of predefined Warnings. The user configurations can now be
imported from an external table, in a normalized format.
The machine and vacuum operation modes can be taken
into account in the generation of a Warning.
Archiving
To avoid saturation of the archives by noisy channels,
“archive smoothing” is automatically selected, with adaptive parameters; the archiving being stopped only if
smoothing is inefficient; additionally, every night all
blocked channels are restarted. After all, if a channel
appears too noisy on a trend, a filter may be used to
extract its minimum values at a given rate.
A large amount of maintenance, consolidation, renovation and optimization works will be performed on the
vacuum control systems, during LS1. Overall, this
involves controllers for more than 1 500 instruments,
modified or manufactured, and then tested, installed &
commissioned. The installation of some 3 000 cables had
to be defined, and later followed, tested & connected.
New Instruments
PS
AD
CPS other
SPS
LHC
R2E LHC
HIE-ISL
L4
NA62
nTOF
TOTAL
372
293
49
74
392
132
142
102
19
1 575
New Cables
365
29
6
229
562
480
390
350
200
60
2 671
Replaced Cables
21
3
276
8
308
Table 3: instrument and cable count for all projects
One of the major challenges will be to have all
machines re-commissioned in time for the restart of the
LHC and injectors, and ready for operation up to 14 TeV.
REFERENCES
[1] J.M. Jimenez et al., “LHC Vacuum Upgrade During
LS1”, Chamonix 2012 Workshop on LHC
Performance, Feb. 2012.
[2] E. Blanco et al., “UNICOS CPC v6: evolution”,
ICALEPCS11, Grenoble, Oct. 2011.
[3] S. Blanchard et al, “Bake-out Mobile Controls for
Large Vacuum Systems”, ICALEPCS13, S.
Francisco, Oct. 2013.
[4] G. Pigny et al, “Mitigation of Radiation and EMI
Effects on the Vacuum Control System of the LHC”,
TWEPP13, Perugia, Sep. 2013.
[5] F. Antoniotti et al, “Quality Management of CERN
Vacuum Controls”,ICALEPCS13, S. Francisco,
Oct. 2013.
[6] F. Antoniotti et al, “Developments on the SCADA of
CERN Accelerators Vacuum”, ICALEPCS13,
S. Francisco, Oct. 2013.